Hydrogen fuel cell hybrid two-wheeler power distribution optimization method based on deep learning

Abstract

The application discloses a hydrogen fuel cell hybrid two-wheeled vehicle power distribution optimization method based on deep learning, which comprises the following steps: step one, collecting vehicle operation data and fuel cell and power battery operation data to construct a short-time state segment; step two, performing energy event segmentation identification according to the short-time state segment to obtain an event segmentation mark sequence; step three, inputting the vehicle operation state and the event segmentation mark sequence into an improved Neural ODE model to obtain a fuel cell slow change reference power trajectory; step four, calculating a power difference value and generating a power battery pulse compensation trajectory and a recovery reserved window trajectory; step five, calculating a driving power gap and performing constraint correction; step six, performing constraint calculation to obtain a target output power; and step seven, performing parameter correction. The application realizes event-driven continuous-time power distribution and parameter adaptive optimization.

Images